User Guide
Programme Insights User Guide
Complete Reference for Platform Users
Version 1.0 | April 2026
Programme Insights | programmeinsights.com
Table of Contents
-
1
Getting Started
3
-
Quick Start: Your First Assessment in 15 Minutes
3
-
Your Dashboard
4
-
User Roles
5
-
2
Connecting Your Documents
6
-
SharePoint Sync
6
-
Manual Upload
8
-
Document Organisation
9
-
3
Running Assessments
11
-
Selecting a Module
11
-
Configuring Your Assessment
12
-
Understanding Assessment Progress
14
-
Batch Assessments
15
-
4
Understanding Your Results
17
-
The Assessment Dashboard
17
-
Criterion-by-Criterion View
18
-
Reading Evidence Citations
20
-
Gap Analysis
21
-
5
Reports
23
-
Generating Reports
23
-
Report Formats
24
-
Sharing Reports
25
-
6
Configuring Custom Criteria
27
-
The Configurable Engine
27
-
Best Practices for Custom Criteria
29
-
7
Administration
31
-
User Management
31
-
Module Configuration
32
-
Data Management
33
-
8
Troubleshooting
34
-
Glossary
37
1. Getting Started
Everything you need to run your first assessment, navigate the dashboard, and understand what you can do with your account.
Quick Start: Your First Assessment in 15 Minutes
You can go from login to assessment findings in under 15 minutes. Here's how.
- Log in to Programme Insights at
app.programmeinsights.com. You'll land on your dashboard.
- Select a module from the dashboard. Each module card shows its name, the criteria framework it assesses against, and the number of criteria. Pick one that matches your current need — for example, IPA Gateway Review for a government programme approaching a gateway.
- Upload your documents. Drag and drop PDF, Word (.docx), or Excel (.xlsx) files into the upload area. Three to five key documents is enough for a first assessment.
- Configure assessment scope. Choose Full Assessment to run against all criteria, or select specific criteria areas to focus on. For your first run, Full Assessment gives you the broadest view.
- Start the assessment. Click Run Assessment. You'll receive an email when it's complete — typically 15–30 minutes for a small document set.
- Review your findings in the assessment dashboard. You'll see an overall RAG rating, criteria coverage percentage, and a prioritised list of findings with evidence citations.
Tip
Start with a small document set for your first assessment — 3–5 key documents. This lets you validate the assessment quality before running a full programme suite.
Your Dashboard
When you log in, the dashboard shows you everything at a glance.
Module Selection
The top of your dashboard displays available assessment modules as cards. Each card shows:
- Module name (e.g., IPA Gateway Review, Green Book Appraisal)
- Number of criteria in the framework
- Brief description of what the module assesses
- Your most recent assessment date and RAG status for that module
Recent Assessments
Below the module cards, a table lists your recent assessments with:
- Assessment name and module
- Date run
- Overall RAG rating
- Criteria coverage percentage
- Quick actions: View Results, Generate Report, Re-run
Quick Actions
From the dashboard you can:
- New Assessment — start a fresh assessment against any module
- Upload Documents — add documents to your library without starting an assessment
- View Reports — access previously generated reports
- Manage Documents — review, tag, and organise your document library
User Roles
Programme Insights has three user roles. Your role determines what you can see and do.
| Role |
Can Do |
Cannot Do |
| Admin |
Everything — run assessments, view findings, generate reports, manage users, configure modules, set up custom criteria, manage SharePoint connections |
— |
| Reviewer |
Run assessments, upload documents, view findings, generate and share reports |
Manage users, configure modules, set up custom criteria |
| Viewer |
View assessment results, view and download reports |
Run assessments, upload documents, manage anything |
Note
Viewers are useful for stakeholders who need to see assessment findings and reports but aren't running assessments themselves — SROs, governance leads, or board members reviewing programme health.
2. Connecting Your Documents
Programme Insights assesses your documents against criteria frameworks. The better organised your documents, the more accurate your findings. There are two ways to connect documents: SharePoint Sync for continuous monitoring, or manual upload for ad-hoc assessments.
SharePoint Sync (Recommended)
SharePoint Sync connects Programme Insights directly to your SharePoint document libraries. Documents are automatically ingested and kept up to date — when a document changes in SharePoint, the platform detects the update and can re-assess automatically.
How to Connect SharePoint
- Navigate to Settings > Document Sources. Click Add SharePoint Connection.
- Authenticate. Sign in with your Microsoft 365 account. Programme Insights requests read-only access to the document libraries you select — it never modifies your SharePoint content.
- Select your SharePoint site from the list of sites you have access to.
- Choose document libraries. Select one or more libraries to sync. You can narrow the scope to specific folders within a library.
- Set sync frequency. Choose how often Programme Insights checks for changes:
- Real-time — checks every 15 minutes (recommended for active programmes)
- Daily — syncs once per day at a time you set
- Manual — syncs only when you trigger it
- Map document types (optional). If your library uses metadata or naming conventions, you can map them to Programme Insights document categories. This improves assessment accuracy.
Benefits of SharePoint Sync
- Continuous monitoring — know when your compliance position changes without running manual assessments
- Automatic re-assessment — when a document is updated in SharePoint, the platform can re-run affected assessments
- No manual upload — your team doesn't need to remember to upload documents before each review
- Version tracking — Programme Insights tracks which version of each document was assessed
SharePoint Troubleshooting
| Issue |
Likely Cause |
How to Fix |
| Sync shows "Pending" |
Permissions haven't propagated yet, or your M365 admin hasn't approved the connection |
Wait 10 minutes. If still pending, ask your M365 admin to check the app consent in Azure AD. |
| Documents not appearing |
Wrong library path, or documents are in a subfolder that wasn't selected |
Go to Settings > Document Sources, edit the connection, and verify the library path. Expand the folder tree to check subfolder selection. |
| Sync errors on specific files |
File is locked, corrupted, or in an unsupported format |
Check the sync log (Settings > Document Sources > Sync History) for the specific file error. Unlock or re-save the file in SharePoint. |
| "Connection expired" |
Microsoft authentication token has expired (typically after 90 days) |
Click Reconnect on the SharePoint connection and re-authenticate with your M365 account. |
Manual Upload
For ad-hoc assessments or when SharePoint isn't available, upload documents directly.
Supported Formats
| Format |
Extension |
Notes |
| PDF |
.pdf |
Must contain searchable text (not scanned images without OCR). See warning below. |
| Microsoft Word |
.docx |
Full support including tables, headings, and tracked changes (resolved version). |
| Microsoft Excel |
.xlsx |
Useful for risk registers, programme schedules, and tabular data. |
Upload Limits
- Per file: 50MB maximum
- Per assessment batch: 500MB maximum
- Per upload session: 50 files at once via drag-and-drop
How to Upload
- Go to Documents > Upload or click Upload Documents from the dashboard.
- Drag and drop files into the upload area, or click Browse Files to select from your file system.
- Review the upload list. Each file shows a progress bar, file size, and format validation status.
- Add tags (optional). Tagging documents by type (e.g., Business Case, Risk Register, Benefits Realisation Plan) helps the platform locate relevant evidence more efficiently.
- Click Upload. Files are processed and added to your document library.
Warning
Scanned PDFs without OCR text layers will have limited assessment coverage. The platform cannot extract text from images. For best results, ensure documents contain searchable text — most modern scanners have an OCR option that adds a text layer automatically.
Document Organisation
How Programme Insights Identifies Documents
The platform analyses document content, headings, and metadata to identify what type of document it is — business case, risk register, programme schedule, and so on. You don't need to organise documents into specific folders for the platform to work.
However, you can improve accuracy by:
- Using clear file names —
Strategic_Outline_Case_v2.1.docx is better than SOC_final_FINAL_v2.docx
- Adding tags when you upload — this gives the platform an explicit signal about document purpose
- Keeping one document per file — a combined document with the OBC and FBC in a single file is harder to assess than two separate files
Managing Document Versions
When you upload a new version of a document, Programme Insights keeps the previous version in your library. You can:
- View which version was used in each assessment
- Re-run a previous assessment against the updated document to see what changed
- Archive old versions to keep your library clean
Tip
Use the
Document Naming Best Practice: include the document type, programme name, and version number in the file name. Example:
HS2_Phase2_FBC_v3.2.docx. This makes documents easy to find and helps the platform identify them accurately.
3. Running Assessments
Assessments are the core of Programme Insights. You select a criteria framework, point it at your documents, and the platform produces a structured, evidence-based assessment with traceable findings.
Selecting a Module
Each module assesses your documents against a specific criteria framework. Choose the module that matches the review, gateway, or compliance requirement you're preparing for.
| Module |
Criteria Count |
When to Use |
| IPA Gateway Review |
218 |
Preparing for an IPA gateway review (Gates 0–5) |
| Green Book Appraisal |
98 |
Assessing a business case against HM Treasury Green Book |
| NEC4 Compliance |
72 |
Checking contract documentation against NEC4 requirements |
| Nuclear (ONR SAPs) |
36 |
Nuclear programme compliance against ONR Safety Assessment Principles |
| CDM Compliance |
48 |
Construction phase health and safety documentation review |
| DCO Compliance |
84 |
Development Consent Order application readiness |
| Custom |
Variable |
Your own criteria framework (see Section 6) |
Note
Not sure which module to use? Start with the one that matches your nearest review milestone. If you're a consultancy working across multiple frameworks, you can run the same documents against different modules to get a comprehensive view.
Configuring Your Assessment
Once you've selected a module, configure how the assessment runs.
Assessment Scope
- Full Assessment — runs against every criterion in the framework. Best for comprehensive reviews and first-time assessments.
- Selected Criteria Areas — choose specific categories or individual criteria to assess. Useful when you know which areas need attention, or when preparing for a focused review.
Custom Criteria
If you've configured custom criteria sets (see Section 6), you can select them here instead of or in addition to the standard module criteria.
Assessment Parameters
| Parameter |
Options |
Recommendation |
| Confidence Threshold |
High, Medium, Low |
Start with Medium. High threshold means the platform only reports findings it's very certain about — fewer findings but higher precision. Low threshold catches more potential issues but includes more that need manual verification. |
| Evidence Depth |
Summary, Standard, Detailed |
Standard for most assessments. Detailed when you need to present findings to a gateway review panel and need full citation trails. |
| Schedule |
Run Now, Schedule for Later |
Schedule large assessments for out-of-hours if you prefer — there's no performance difference, but it avoids the "waiting" feeling during your working day. |
Starting the Assessment
- Review your configuration — the summary panel shows: module, criteria count, document count, and estimated completion time.
- Click Run Assessment.
- You'll see a confirmation with the assessment ID and estimated completion time. An email notification is sent when findings are ready.
Understanding Assessment Progress
While assessments run, the platform streams real-time progress updates directly to your browser using SSE (Server-Sent Events). You can watch each criterion being assessed live — for example, "Assessing criterion 12 of 62: Client-Side Budget" — or close the browser and come back later. An email notification is sent when the assessment completes.
Assessment Tiers
When configuring an assessment, you choose the assessment depth tier. This determines how thoroughly each criterion is analysed.
| Tier |
What It Does |
Best For |
| Standard |
Runs two agentic steps per criterion: hybrid retrieval (combining semantic and keyword search) followed by scoring with deterministic guardrails. Timeout: 300 seconds per criterion. |
Regular assessments, progress tracking, initial gap identification. Fast and reliable. |
| Extended |
Adds Corrective RAG (re-retrieval for ambiguous results), citation verification, self-critique for low-confidence findings, and adversarial debate for boundary cases. Timeout: 480 seconds per criterion. |
Pre-gateway assessments, board-ready findings, or when you need the highest confidence in results. |
Tip
Start with Standard tier for your first assessment to get a quick baseline. Switch to Extended tier when preparing for a formal review or gateway — the additional scrutiny catches edge cases that matter when findings are presented to a panel.
Status Indicators
Ingesting
→
Assessing
→
Finalising
→
Complete
| Status |
What's Happening |
Typical Duration |
| Ingesting |
Documents are being parsed — text extraction, structure analysis, chunking, and embedding for retrieval |
2–5 minutes |
| Assessing |
Each criterion is assessed individually: retrieval, sub-question scoring, and (in Extended tier) corrective re-retrieval, citation verification, self-critique, and adversarial debate. Real-time progress shows which criterion is being processed. |
Depends on criteria count and tier. A 62-criterion module on Standard tier: ~20 minutes. On Extended tier: ~35 minutes. |
| Finalising |
Dimension-level RAG ratings are calculated, showstoppers checked, and the overall DCA (Delivery Confidence Assessment) is produced combining rules-based analysis with holistic professional judgement. |
1–3 minutes |
| Complete |
Findings are ready for review. Email notification sent. |
— |
Carry-Forward: Faster Subsequent Assessments
When you re-run an assessment against the same module, Programme Insights uses carry-forward to skip criteria that aren't affected by document changes. The platform tracks evidence lineage — which document chunks were used for each criterion — and only re-assesses criteria where the underlying documents have been added, removed, or modified. This makes subsequent assessments significantly faster than the first run.
Note
Overall assessment time depends on criteria count, document volume, and the tier you select. Carry-forward reduces time further for repeat assessments. A first-time full IPA assessment (218 criteria, Extended tier) may take 60–90 minutes; a re-run after updating a few documents typically completes in under 15 minutes.
Batch Assessments
Batch assessments let you run the same criteria framework across multiple document sets in a single operation. This is useful for:
- Consultancies reviewing multiple client programmes against the same framework
- Portfolio managers comparing compliance across programmes in a portfolio
- Trend analysis — running the same assessment against different versions of your documents over time to track improvement
How to Run a Batch Assessment
- Go to Assessments > New Batch Assessment.
- Select a module — all assessments in the batch use the same criteria framework.
- Add document sets. Each document set represents one programme or submission. Name each set clearly (e.g., "Programme A — Gate 3", "Programme B — Gate 3").
- Configure parameters (same options as a single assessment — these apply to all sets in the batch).
- Run Batch. Each document set is assessed independently. Results are presented in a comparison view when all assessments complete.
Batch Results: Comparison View
When a batch completes, you get a comparison dashboard showing:
- Side-by-side RAG ratings per document set
- Criteria coverage comparison
- Common gaps across all document sets (issues appearing everywhere)
- Unique gaps (issues specific to one programme)
Tip
Use batch assessments for trend analysis: run the same criteria against your documents monthly. The comparison view shows which gaps have been addressed and which remain open — a clear improvement trajectory to present at governance boards.
4. Understanding Your Results
Assessment results are where Programme Insights delivers its value. Every finding is traceable to evidence in your documents, every score is deterministic, and every gap comes with actionable context.
The Assessment Dashboard
When an assessment completes, the dashboard gives you an at-a-glance view of your position.
Overall DCA Rating
The assessment produces an overall Delivery Confidence Assessment (DCA) rating. This is not a simple average — it combines rules-based dimension analysis with holistic professional judgement. Criteria are grouped into dimensions, per-dimension RAG ratings are calculated, showstoppers are checked, and the final DCA reflects a considered overall position:
| Rating |
Meaning |
Action |
| Green |
Strong evidence found for the majority of criteria. Few or no significant gaps. |
Spot-check findings. Prepare for review with confidence. |
| Amber |
Evidence found for most criteria, but some gaps or incomplete evidence in key areas. |
Review amber and red criteria. Address gaps before the review. |
| Red |
Significant gaps in evidence. Multiple criteria not met or insufficient evidence. |
Prioritise gap remediation. Consider whether the programme is ready for review. |
Criteria Coverage
Criteria coverage tells you what percentage of criteria had evidence found in your documents. A 100% coverage doesn't mean everything passed — it means the platform found relevant evidence (positive or negative) for every criterion. Low coverage indicates documents that don't address the full criteria framework.
Executive Summary
The top of the assessment dashboard shows a generated executive summary with four components:
- Verdict headline — a one-sentence summary of the overall assessment position
- Key strengths — areas where evidence is strong and criteria are well met
- Critical issues — RED-rated criteria and showstoppers that need immediate attention
- Overall recommendation — a summary judgement on readiness, with suggested next steps
Criterion-by-Criterion View
Click any criterion to see its full assessment detail.
What You See for Each Criterion
Each criterion now produces two distinct ratings:
- Coverage Rating — how well the evidence in your documents covers what the criterion requires (based on sub-question scores)
- Delivery Confidence Rating — the platform's confidence in the quality and sufficiency of that evidence for a real review
Example: IPA Criterion 4.2.1
"The programme has a quantified benefits realisation plan with baseline measurements and target outcomes."
| Coverage Rating |
Amber (58%) — partial evidence found across sub-questions |
| Delivery Confidence |
Amber — evidence exists but gaps in baseline data would be flagged in a review |
| Confidence |
High (85%) |
| Evidence |
Benefits_Realisation_Plan_v2.docx, page 14: "Target outcomes have been identified for each workstream..."
Strategic_Outline_Case.pdf, page 42: "Baseline measurements will be established during..." |
| Finding |
Benefits realisation plan exists with target outcomes defined. However, baseline measurements are described as future activity ("will be established") rather than completed. Partial compliance. |
Note
When the platform's confidence in a finding drops below 70%, it automatically triggers a self-critique step — re-examining the evidence and reasoning before finalising the rating. In Extended tier, AMBER-rated criteria also undergo adversarial debate, where multiple AI perspectives challenge the finding to ensure it holds up under scrutiny.
Sub-Question Breakdown
Programme Insights decomposes each criterion into specific sub-questions to produce traceable, deterministic scores. Each sub-question is scored as Yes (100%), Partial (70%), or No (0%). The coverage percentage is calculated from these scores and maps to a RAG rating: GREEN at 70% or above, AMBER at 40–69%, RED below 40%.
For the example above:
| Sub-Question |
Score |
Evidence |
| Does a benefits realisation plan exist? |
Yes (100%) |
Benefits_Realisation_Plan_v2.docx found in document set |
| Are benefits quantified with measurable targets? |
Yes (100%) |
Page 14: target outcomes per workstream with KPIs |
| Are baseline measurements documented? |
No (0%) |
Page 42 of SOC references future baseline activity; no current baselines found |
| Is there a schedule for benefits tracking? |
Partial (70%) |
Page 18: quarterly benefits review cycle defined, but no specific dates or milestones |
In this example, the coverage score is (100 + 100 + 0 + 70) / 4 = 67.5%, which maps to AMBER. Deterministic guardrails then apply: these rules-based overrides ensure the RAG rating is consistent with the sub-question evidence, preventing the LLM from rating a criterion GREEN when the sub-question scores indicate otherwise.
This decomposition means every score is traceable: you can see exactly which sub-questions produced the overall criterion rating and verify each one against the source evidence.
Reading Evidence Citations
Every finding in Programme Insights is backed by evidence citations. Understanding how to read and verify them is essential.
What a Citation Looks Like
Each citation includes:
- Document name — the source file
- Page or section reference — where in the document
- Text excerpt — the relevant passage (typically 1–3 sentences)
- Confidence level — how certain the platform is that this evidence is relevant to the criterion
Confidence Levels
| Level |
What It Means |
Your Action |
| High (80%+) |
Strong, direct evidence. The cited text explicitly addresses the criterion. |
Spot-check a sample. These findings are reliable. |
| Medium (50–79%) |
Relevant evidence found, but it may be indirect or partially addresses the criterion. |
Review the citation in context. Determine whether it's sufficient or needs strengthening. |
| Low (<50%) |
Possible evidence, but the platform isn't certain about relevance. May be tangential. |
Verify manually. This evidence may not hold up in a review. |
Tip
Start your review with RED-rated criteria — these represent the highest-risk gaps. Then review AMBER criteria where evidence was found but incomplete. GREEN criteria have strong evidence and typically need only spot-checking.
Verifying a Citation
- Click the citation link in the findings view. This opens the source document at the referenced page.
- Read the cited text in context. The excerpt is highlighted. Check that the surrounding paragraphs support the finding.
- Confirm or flag. Use the thumbs-up/thumbs-down control to confirm the citation is accurate or flag it for review. Flagged citations are excluded from report generation until resolved.
Gap Analysis
A gap is a criterion where Programme Insights found no sufficient evidence in your document set. Gaps are the most actionable output of an assessment — they tell you exactly where your documentation needs work before a review.
The Gap Register
The Gap Analysis view presents all gaps as a prioritised register:
| Column |
What It Shows |
| Criterion |
The criteria reference and description |
| Category |
Which area of the framework (e.g., "Benefits Management", "Risk") |
| Priority |
High / Medium / Low based on criterion weight and potential review impact |
| Recommendation |
What to do about the gap — which document to create or update, what content to add |
| Owner |
Assignable field for tracking who's addressing the gap |
| Status |
Open / In Progress / Resolved |
What Makes a Gap High Priority?
Gap priority is determined by:
- Criterion weight — some criteria carry more weight in the framework (e.g., IPA "deliverability" criteria are heavily weighted)
- Review impact — gaps that are commonly flagged in reviews receive higher priority
- Related gaps — if multiple criteria in the same area are missing evidence, the area as a whole is flagged
Acting on Gaps
Each gap includes a recommendation with specific guidance:
- Which document should contain this evidence
- What type of content is needed (e.g., "quantified baseline measurements", "risk mitigation actions with owners and dates")
- Whether the gap is likely a documentation gap (evidence exists but isn't documented) or a genuine programme gap (the work hasn't been done)
Warning
A documentation gap and a programme gap require different responses. A documentation gap means updating a document. A programme gap means the programme needs to do actual work — commission a study, complete an analysis, hold a stakeholder engagement. Don't paper over genuine programme gaps with better documentation.
5. Reports
Programme Insights generates board-ready reports with traceable evidence. Every finding in a report links back to source documents, making the output suitable for governance boards, gateway preparation packs, and SRO briefings.
Generating Reports
From any completed assessment, click Generate Report to create a formatted output. You have four report types available.
Report Types
| Report |
Audience |
Content |
Length |
| Assessment Report |
Programme team, reviewers, PMO |
Full criterion-by-criterion findings with evidence citations, sub-question breakdown, and recommendations |
20–60 pages depending on criteria count |
| Executive Summary |
SRO, board, senior stakeholders |
One-page overview: overall RAG, key findings, critical gaps, recommended actions |
1–2 pages |
| Gap Analysis Report |
Programme team, PMO, consultancy leads |
Prioritised gap register with recommendations, grouped by category, with improvement trajectory if trend data is available |
5–15 pages |
| Custom Report |
Any |
Select which sections to include: pick specific criteria areas, include or exclude sub-questions, choose evidence detail level |
Variable |
How to Generate
- Open the completed assessment from your dashboard or the Assessments list.
- Click Generate Report in the top-right action bar.
- Select report type from the dropdown.
- Configure options (for Custom Reports): select criteria areas, evidence depth, and whether to include trend data from previous assessments.
- Click Generate. Report generation takes 1–2 minutes. You'll see a download link when it's ready.
Report Formats
PDF Output
All reports are generated as professionally formatted PDF documents with:
- Branded cover page with assessment details, date, and module name
- Table of contents with clickable section links
- Consistent formatting — RAG indicators, evidence callouts, and structured tables throughout
- Traceable citations — every finding references the source document and page
- Print-ready — A4 layout designed for printing or sharing as-is to governance boards
Tip
The Executive Summary report is specifically designed to be handed to an SRO or board member without additional context. It's self-contained: the reader doesn't need to have seen the assessment dashboard to understand the findings.
Data Export
In addition to PDF reports, you can export raw assessment data for integration with other tools:
- Excel export — full criteria results in a spreadsheet, useful for importing into your own tracking systems
- CSV export — gap register data for import into project management tools
- JSON API — programmatic access for organisations integrating Programme Insights with their own dashboards or reporting tools
Sharing Reports
Download
Click the download icon on any generated report to save it locally as a PDF. Reports are also stored in your Reports library for later access.
Share via Link
Generate a shareable link for any report. Recipients don't need a Programme Insights account to view the report — but you control access:
- Anyone with link — suitable for sharing with external stakeholders (link expires after 30 days)
- Specific people — enter email addresses; recipients must verify their identity to view
- Organisation only — only users with accounts in your organisation can access
Revoking Access
Shared links can be revoked at any time from Reports > Shared Links. Once revoked, the link immediately stops working.
Note
Reports contain assessment findings that may be commercially sensitive. Use the "Specific people" or "Organisation only" sharing options for reports that include programme-level detail. The "Anyone with link" option is best suited for summary-level reports.
6. Configuring Custom Criteria
Programme Insights isn't locked to predefined frameworks. The configurable engine lets you add your own criteria sets — particularly useful for consultancies with proprietary methodologies, organisations with internal governance frameworks, or programmes with bespoke requirements.
The Configurable Engine
Every standard module (IPA, Green Book, NEC4, etc.) uses the same underlying engine. Custom criteria work identically: you define the criteria, the platform decomposes them into sub-questions, matches against your documents, and produces traceable, evidence-based findings.
How to Add a Custom Criteria Set
- Go to Settings > Criteria Management > Create New Set.
- Name your criteria set. Use something descriptive — e.g., "Mott MacDonald P3M3 Framework" or "Internal Programme Governance Checklist".
- Define categories. Group related criteria into categories (e.g., "Governance", "Benefits Management", "Stakeholder Engagement"). This determines how findings are organised in reports.
- Add criteria. For each criterion, provide:
- Reference — a unique identifier (e.g., "GOV-01", "BEN-03")
- Description — what the criterion requires (see best practices below)
- Category — which category it belongs to
- Weight (optional) — relative importance for overall scoring
- Evidence description (optional) — what type of evidence would demonstrate compliance
- Review sub-questions. The platform automatically generates binary decomposition sub-questions for each criterion. Review them and adjust if needed — these determine exactly how the criterion is assessed.
- Test your criteria set. Run a test assessment against a small document set to verify the criteria produce meaningful findings. Make adjustments before publishing.
- Publish. Move the criteria set from Draft to Active. It now appears as an option when creating assessments.
Note
Draft criteria sets are only visible to the user who created them. Once published (Active), they're available to all Reviewers and Admins in your organisation.
Binary Decomposition: How to Write Effective Sub-Questions
The platform generates sub-questions automatically, but reviewing and refining them produces better results. Effective sub-questions are:
- Binary — answerable as yes or no, not "somewhat" or "partially"
- Specific — reference concrete evidence (documents, data, processes)
- Independent — each sub-question stands alone; the answer to one shouldn't depend on another
- Observable — the answer can be determined from document evidence, not from interviews or tacit knowledge
Good Sub-Questions
"Does the benefits realisation plan include quantified baseline measurements?"
"Is there a named benefits owner for each identified benefit?"
"Does the plan include a schedule for benefits tracking and review?"
Poor Sub-Questions
"Is the benefits plan comprehensive?" (too vague — what counts as comprehensive?)
"Are stakeholders satisfied with the benefits approach?" (can't determine from documents)
"Does the plan adequately address benefits?" (subjective — who defines 'adequately'?)
Best Practices for Custom Criteria
Writing Effective Criteria
| Principle |
Example |
| Be specific and testable |
"The programme includes a quantified benefits realisation plan with baseline measurements" rather than "The programme demonstrates good benefits management" |
| Include clear evidence descriptions |
"Evidence: a benefits register with named owners, target values, baseline measurements, and a tracking schedule" — this tells the platform exactly what to look for |
| Group logically |
Related criteria in the same category produce a coherent section in reports. "Risk Management" criteria should all live in a "Risk" category, not scattered across "Governance" and "Delivery". |
| Start small, then expand |
Begin with your top 20 most important criteria. Run a test assessment. Refine. Then add the next 20. |
Warning
Custom criteria quality directly affects assessment quality. Vague criteria like "demonstrates good project management" produce vague findings. Specific criteria like "programme includes a quantified benefits realisation plan with baseline measurements" produce actionable findings.
Managing Criteria Sets
- Versioning — when you update criteria, the platform creates a new version. Previous assessments remain linked to the criteria version they were run against.
- Sharing — Active criteria sets can be shared across your organisation. For consultancies, you can create criteria sets that are available to specific teams or projects.
- Importing — if you have criteria in a spreadsheet, you can import via Excel upload (Settings > Criteria Management > Import). The required columns are: Reference, Description, Category.
Tip
If you're a consultancy with a proprietary assessment methodology, your criteria set is your intellectual property implemented in software. Every assessment you run using custom criteria produces consistent, traceable results — the same assessment quality regardless of which team member runs it.
7. Administration
Administration functions are available to users with the Admin role. This section covers user management, module configuration, and data management.
User Management
Adding Users
- Go to Settings > Users > Invite User.
- Enter the user's email address.
- Select a role: Admin, Reviewer, or Viewer (see Section 1 for role descriptions).
- Click Send Invitation. The user receives an email with a link to set up their account.
Invitations expire after 7 days. If a user hasn't accepted, you can resend from the Users list.
Changing Roles
Go to Settings > Users, find the user, and change their role from the dropdown. The change takes effect immediately — the user doesn't need to log out and back in.
Removing Access
Click Remove next to the user in the Users list. Their access is revoked immediately. Assessment results they created are not deleted — they remain in the system linked to their name for audit trail purposes.
Note
At least one Admin must exist in every organisation. You cannot remove the last Admin — assign Admin to another user first.
Module Configuration
Enabling and Disabling Modules
Not every organisation needs every module. Go to Settings > Modules to enable or disable assessment modules. Disabled modules don't appear on the dashboard or in the module selection when creating assessments.
Default Assessment Parameters
Set organisation-wide defaults for assessment configuration:
- Default confidence threshold — applied to all new assessments unless overridden
- Default evidence depth — Summary, Standard, or Detailed
- Email notifications — who receives assessment completion notifications (creator only, or all team members)
Custom Criteria Sets
Admins can manage all custom criteria sets in the organisation — including those created by other users. Go to Settings > Criteria Management to view, edit, publish, or archive criteria sets.
Data Management
Assessment Data Retention
By default, all assessment data is retained indefinitely. You can configure retention policies in Settings > Data Management:
- Keep all data (default) — assessments, documents, and reports are retained until manually deleted
- Auto-archive after N months — assessments older than the specified period are archived (still accessible but moved to cold storage)
- Auto-delete after N months — assessments and associated documents are permanently deleted after the specified period
Warning
Auto-delete is permanent. Deleted assessments, documents, and reports cannot be recovered. If your organisation has regulatory requirements for data retention, ensure the retention period meets those requirements before enabling auto-delete.
Exporting Assessment History
Export your full assessment history as an Excel spreadsheet from Settings > Data Management > Export History. The export includes:
- All assessment metadata (date, module, user, document count)
- Overall RAG ratings and criteria coverage scores
- Gap counts per category
This is useful for governance reporting, demonstrating improvement trends to boards, or migrating data if needed.
Backup and Recovery
Programme Insights automatically backs up all data daily. In the event of data loss:
- Contact support at
support@programmeinsights.com
- Provide your organisation ID (found in Settings > Organisation)
- Specify the date range of the data you need recovered
Recovery typically takes 2–4 hours during business hours.
8. Troubleshooting
Common issues and how to resolve them. For issues not covered here, contact support at support@programmeinsights.com.
Assessment Issues
| Issue |
Likely Cause |
How to Fix |
| Assessment stuck at "Ingesting" |
One or more documents have a format issue — typically a scanned PDF without an OCR text layer, or a corrupted file. |
Check the assessment log (click the assessment status for details). The log identifies which document caused the issue. Remove or replace the problem file and re-run. |
| Assessment stuck at "Analysing" for over 2 hours |
Very large document set (>100 documents) or a system processing delay. |
Check the assessment log for progress indicators. If no progress in 2+ hours, contact support with your assessment ID. |
| Low coverage score despite complete documents |
Documents may not contain the right headings or structure for the platform to identify relevant sections. Or documents were uploaded in an image format without OCR. |
Check document naming and add tags. Re-upload any scanned documents with OCR applied. For very unstructured documents, add document type tags manually. |
| Unexpected RED ratings on criteria you know are covered |
Evidence may exist in a document the platform couldn't parse fully, or the evidence uses different terminology than the criterion expects. |
Check the sub-question breakdown for the criterion. If a sub-question is marked RED but you know the evidence exists, check which document should contain it and verify the document was uploaded and parsed correctly. |
Document Issues
| Issue |
Likely Cause |
How to Fix |
| Upload fails |
File exceeds 50MB limit, unsupported format, or network timeout. |
Check file size and format. For large files, try uploading individually rather than in a batch. If the format is correct and under 50MB, try again — intermittent network issues can cause upload failures. |
| Citation references wrong page |
The document may have been modified after assessment. Page numbers shift when content is added or removed. |
Re-run the assessment against the current version of the document. The new results will reference the correct pages. |
| Document shows as "unsupported format" |
File is a legacy format (.doc instead of .docx), password-protected, or has DRM restrictions. |
Save the file as .docx or .pdf without password protection. Programme Insights cannot open password-protected or DRM-restricted files. |
SharePoint Issues
| Issue |
Likely Cause |
How to Fix |
| SharePoint sync not updating |
Authentication token expired, or sync frequency set to Manual. |
Go to Settings > Document Sources. Check connection status. If expired, click Reconnect. Check sync frequency is set to Real-time or Daily. |
| "Permission denied" on specific files |
Your M365 account doesn't have read access to those files or folders. |
Ask your SharePoint admin to grant read access to the affected files. Programme Insights can only sync files your account has permission to read. |
Report Issues
| Issue |
Likely Cause |
How to Fix |
| Cannot generate report |
Assessment may still be running, or the assessment completed with errors. |
Check assessment status. If it's still running, wait for completion. If it shows errors, check the assessment log and address the issues before generating a report. |
| Shared link not working for recipients |
Link has expired (30-day default for "anyone with link"), or recipient's email isn't on the specific people list. |
Generate a new link from Reports > Shared Links. Verify the recipient's email address if using "Specific people" access. |
Custom Criteria Issues
| Issue |
Likely Cause |
How to Fix |
| Custom criteria not appearing in module selection |
Criteria set is still in Draft status. |
Go to Settings > Criteria Management. Find your set and click Publish to move it from Draft to Active. |
| Criteria producing poor-quality findings |
Criteria descriptions are too vague, or sub-questions need refinement. |
Edit the criteria to be more specific. Review and adjust the auto-generated sub-questions. Run a test assessment after changes to verify improvement. |
| Excel import fails |
Missing required columns (Reference, Description, Category) or formatting issues. |
Download the import template from Settings > Criteria Management > Import. Use the template to format your data correctly, then re-import. |
Account Issues
| Issue |
Likely Cause |
How to Fix |
| Can't log in |
Wrong password, account locked after failed attempts, or invitation not yet accepted. |
Use "Forgot password" on the login page. If your account is locked, wait 30 minutes or ask your Admin to unlock it. If you haven't received an invitation, ask your Admin to resend. |
| Missing features or menu items |
Your user role doesn't have permission for those features. |
Check your role in your profile (top-right menu > My Account). If you need additional access, ask your Admin to change your role. |
Tip
If you're reporting an issue to support, include: your organisation ID (Settings > Organisation), the assessment ID (if relevant), and a screenshot of the error. This helps the support team resolve your issue faster.
Glossary
Key terms used throughout this guide and the Programme Insights platform.
| Term |
Definition |
| Adversarial Debate |
A multi-persona review process triggered for AMBER-rated criteria in Extended tier assessments. Multiple AI perspectives challenge the finding from different angles to ensure the rating holds up under scrutiny before it is finalised. |
| ALARP |
As Low As Reasonably Practicable. A nuclear safety principle requiring that risks are reduced to the lowest level that is reasonably achievable, taking into account cost and practicality. |
| Binary Decomposition |
The process of breaking a complex assessment criterion into specific yes/no sub-questions. This produces deterministic scoring — each sub-question has a clear, evidence-based answer. |
| Carry-Forward |
An optimisation that copies valid results from previous assessment runs and only re-assesses criteria where the underlying documents have been added, removed, or modified. Makes subsequent assessments significantly faster than the first run. |
| CDM |
Construction (Design and Management) Regulations 2015. UK legislation setting out the duties of those involved in construction work to plan, manage, and coordinate health and safety. |
| Citation |
A reference to a specific location in a source document where evidence was found. Includes document name, page number, and text excerpt. |
| Citation Verification |
The automated process of confirming that evidence references in assessment findings accurately match the content in source documents. |
| Corrective RAG (CRAG) |
An Extended tier step that re-retrieves evidence using decomposed sub-queries when initial retrieval results are poor or ambiguous. Improves evidence coverage for complex or multi-faceted criteria. |
| Confidence Level |
A measure of how certain the assessment is about a finding. High (80%+), Medium (50–79%), or Low (<50%). Based on evidence quality, relevance, and directness. |
| Criteria Coverage |
The percentage of assessment criteria for which evidence (positive or negative) was found in the document set. 100% coverage means every criterion was addressed — not that every criterion passed. |
| DCA |
Delivery Confidence Assessment. The overall holistic rating produced for an assessment, combining rules-based dimension analysis with professional judgement. Not a simple average — it accounts for dimension weightings, showstoppers, and cross-cutting themes. |
| DCO |
Development Consent Order. The consent process for nationally significant infrastructure projects in England and Wales under the Planning Act 2008. |
| Deterministic Guardrails |
Rules-based overrides that calculate coverage from sub-question scores and can override LLM ratings. Ensures the RAG rating is consistent with the evidence — for example, preventing a GREEN rating when sub-question scores indicate AMBER. |
| Deterministic Scoring |
Scoring based on evidence found in documents, not probabilistic AI estimation. The same documents assessed against the same criteria produce the same score every time. |
| Evidence Lineage |
The traceable connection between assessment findings and the specific document chunks that produced them. Enables carry-forward by identifying exactly which criteria are affected when documents change. |
| Extended Tier |
The deeper assessment tier that adds Corrective RAG, citation verification, self-critique for low-confidence findings, and adversarial debate for AMBER-rated criteria. Timeout: 480 seconds per criterion. Best for pre-gateway or board-ready assessments. |
| FBC |
Full Business Case. The final stage of business case development in the Five Case Model, produced after procurement and before full programme delivery. |
| Five Case Model |
HM Treasury framework for business case development. The five cases are: Strategic, Economic, Commercial, Financial, and Management. |
| Gap |
A criterion where no sufficient evidence was found in the document set. Gaps may indicate missing documentation or genuine programme gaps requiring action. |
| Green Book |
HM Treasury guidance on appraisal and evaluation of policies, programmes, and projects. The standard framework for government spending proposals. |
| IPA |
Infrastructure and Projects Authority. The UK government's centre of expertise for infrastructure and major project delivery, responsible for the Gateway Review process. |
| NEC4 |
New Engineering and Construction Contract, 4th edition. A family of contracts used extensively in UK construction and infrastructure, designed around collaborative project management principles. |
| OBC |
Outline Business Case. The second stage of business case development in the Five Case Model, produced after the Strategic Outline Case and before procurement. |
| ONR |
Office for Nuclear Regulation. The UK's independent nuclear regulator, responsible for ensuring nuclear safety and security. |
| P3M3 |
Portfolio, Programme and Project Management Maturity Model. A framework for assessing organisational maturity in managing portfolios, programmes, and projects. |
| RAG Rating |
Red/Amber/Green status indicator. Red = significant gaps or non-compliance. Amber = partial evidence or areas of concern. Green = strong evidence, criteria met. |
| Reciprocal Rank Fusion (RRF) |
A technique that combines results from semantic search and keyword search into a single ranked list. Ensures both conceptually relevant and exact-term matches are surfaced during evidence retrieval. |
| SAPs |
Safety Assessment Principles. ONR's framework of principles used to assess the safety of nuclear facilities and operations. |
| SOC |
Strategic Outline Case. The first stage of business case development in the Five Case Model, establishing the strategic need for the programme. |
| Standard Tier |
The default assessment tier that runs two agentic steps per criterion: hybrid retrieval (combining semantic and keyword search via Reciprocal Rank Fusion) followed by scoring with deterministic guardrails. Timeout: 300 seconds per criterion. Fast and reliable for regular assessments. |
| SRO |
Senior Responsible Owner. The individual accountable for ensuring a programme meets its objectives and delivers its benefits. The key decision-maker in programme governance. |
| Sub-question |
A component of a decomposed criterion. Each sub-question is assessable as yes/no based on document evidence. The collection of sub-question answers determines the criterion's RAG rating. |
Programme Insights | programmeinsights.com
User Guide Version 1.0 | April 2026
For support, contact support@programmeinsights.com